| Summary: | 碩士 === 朝陽科技大學 === 建築及都市設計研究所 === 91 === Abstract
A model of dynamic school facility location choice is proposed in this thesis research. A dynamic p-median problem of school facility is considered. One important characteristic of this research is that the school students demand predicted in our model is changing over a given time horizon, i.e., the demand is time-dependent. Besides, the school facilities are built one at a time at given times. Once a new school facility is built, some of the students will use its services and some of them will patronize an existing facility. At any given time, students patronize the closest school facility under enough capacity. The problem is to find the best school facility location for the new facilities. The problem is formulated and the potential algorithm is proposed for this present time. However, a standard mathematical programming code AMPL (A Modeling Language for Mathematical Programming) is adopted here to solving our formulated problem. In this research it is assumed that new facilities are added to the area in given future points in time. The construction of schools, hospitals, public parking lots, stores, restaurants and other facilities in a growing area is typical for such applications.
It is assumed in this research that the demand at each demand point may be time-dependent with a given functional relationship over the time horizon. Several school facilities need to be located over the time horizon. The time when each individual school facility is constructed is given for each new facility. No relocation of school facilities is allowed. This is a reasonable assumption when the relocation cost is relatively high. The question is where to locate each of the school facilities (each school facility is located at a prespecified time) to minimize the total transportation cost over the time horizon. We show that this problem is translated into what we term the progressive p-median problem. The progressive p-median problem is the successive location of facilities each with a different set of demands to minimize the total cost in the system. Finally, a real-world case study located in Pai-toun district of Taichung City in Taiwan will be proposed in this research to evaluate the performance of the research algorithm.
|
|---|
